KR20130136922A - Mask data generation method - Google Patents

Abstract

A method for generating mask data comprises a step for obtaining data of a pattern including a plurality of pattern components; a step for dividing an area of the pattern into a plurality of sections by using the data of the pattern in order that each pattern component is arranged on each section, and generating map data including information which shows the existence or the absence of a pattern component in each section; a step for setting a set of mask entity information per section including the pattern component by using a limitation condition and the map data; and a step for generating the data for masks corresponding to the mask entity information. [Reference numerals] (AA) Start;(BB) Yes;(CC) No;(DD) End;(S101) Obtain pattern data;(S102) Convert the data into map data;(S103) Set a limitation condition;(S104) Select a square;(S105) Obtain predetermined mask entity information in the limitation area;(S106) Set the mask entity information for the square;(S107) Exist the square excluding the mask entity information

Description

How to generate mask data {MASK DATA GENERATION METHOD}

The present invention relates to a method of generating mask data.

An exposure apparatus (lithography apparatus) is used in the lithography process of the manufacturing process of a semiconductor device. A lithography process is a process of transferring the circuit pattern of a semiconductor device onto a board | substrate (called a silicon substrate, a glass substrate, or a wafer). The illumination optical system of the exposure apparatus illuminates the mask (reticle) using the light beam emitted from the light source. Thereafter, the circuit pattern formed on the mask is transferred onto the wafer via, for example, a projection optical system.

Recently, due to the pattern miniaturization of semiconductor devices, a method called multiple patterning is used. According to this multiple patterning, one layer of the wafer is exposed a plurality of times using a plurality of masks. As a result, a plurality of mask patterns are formed on the wafer.

The resolution limit of the exposure apparatus is expressed by hp = k1 × λ / NA. In this equation, "hp" is an abbreviation for half pitch and is half of the shortest distance between two neighboring patterns. Also, "k1" is a process related factor, "λ" is an exposure wavelength, and "NA" is a numerical aperture of the exposure apparatus. According to the multi-patterning, the pattern having a half pitch smaller than the half pitch of the resolution limit of the exposure apparatus is divided into the patterns of the plurality of masks. When the wafer is exposed using the obtained mask pattern, a pattern finer than the resolution limit can be formed on the wafer.

The method of dividing a layout (pattern) into a pattern of a plurality of masks is generally called a coloring problem because it is similar to the coloring of maps. In the following description, the method of dividing a pattern can be described using an expression used for coloring. The method of dividing the original target pattern into the patterns of the plurality of masks is U.S. Patent Application Publication No. 2007/0031740 and U.S. Pat. It is discussed in Patent Application Publication No. 2011/0078638.

U.S.A. Patent Application Publication No. 2007/0031740 describes a method of repeatedly applying a division rule. According to this technique, a division rule is determined. Then, the target pattern is divided into patterns to be formed by the first mask or the second mask based on the division rule. This process is repeated for each pattern.

U.S.A. Patent Application Publication No. 2011/0078638 discusses a method of pattern segmentation using conflict graphs and mathematical programming. The conflict graph consists of vertices and edges. When the conflict graph is applied to pattern division, each mask pattern is represented by a point, and patterns exceeding a resolution limit are connected by sides. Also, the pattern is divided such that the mask numbers of the two regions sharing the border (edge) are different. This process uses integer programming.

On the other hand, with the drive to lower k1 values, it has become difficult to faithfully transfer the desired pattern onto the wafer using conventional two-dimensional layout patterns extending in the X and Y directions. Thus, recently, "Low k1 Logic Design using Gridded Design Rules", Michael C. Smayling, et al., Proc. of SPIE Vol. A method of fabricating circuit patterns called one-dimensional layout techniques, such as the technique discussed in 6925 (2008), has been developed. According to the one-dimensional layout technique, a single pitch line and space (L / S) pattern is formed. Thereafter, at a plurality of positions, a plurality of pattern elements of a hole pattern or a cut pattern having the same size are exposed on regular grids. A single pitch L / S pattern is then cut from the pattern elements and a circuit pattern is produced. According to this method, not only can the exposure area be reduced as compared with the conventional two-dimensional pattern, but the exposure process can be performed more easily.

U.S.A. Patent Application Publication No. 2007/0031740 and U.S. Pat. According to Patent Application Publication No. 2011/0078638, pattern division is performed based on a two-dimensional layout pattern extending in the X and Y directions. U.S.A. According to Patent Application Publication No. 2007/0031740, it is necessary to determine, for both two-dimensional layout patterns, whether the target pattern and other patterns will be applied to the division rule. Accordingly, a long time is required for the calculation. Also, U.S. Pat. According to Patent Application Publication No. 2011/0078638, in order to generate a conflict graph, it is necessary to calculate the distance between patterns for all two-dimensional layout patterns. Therefore, a long calculation time is required.

Also, U.S. Pat. Patent Application Publication No. 2007/0031740 and U.S. Pat. There is no conventional case where the pattern dividing method discussed in Patent Application Publication No. 2011/0078638 is applied to a plurality of pattern elements of a hole pattern or a cut pattern of a one-dimensional layout. Even if the pattern division method is actually applied to a plurality of pattern elements, it is necessary to calculate the distance between all the pattern elements using the data of the plurality of pattern elements, and determine whether the obtained distance satisfies the distance constraint. . Accordingly, a long calculation time is required.

The present invention relates to a mask data generation method useful for generating mask data by dividing a pattern comprising a plurality of pattern elements within a shorter time.

According to an aspect of the present invention, a plurality of masks including the first mask and the second mask used for multiple patterning of exposing the substrate using a second mask and then exposing the substrate using a second mask The mask data generation method for generating data of a computer using a computer includes obtaining data of a pattern including a plurality of pattern elements, and placing each pattern element in each section using the acquired data of the pattern. Dividing the region of the pattern into a plurality of sections so as to generate map data including information indicating the presence or absence of the pattern element in each section, the constraint region including one section and sections around it The pattern for inhibiting setting of the same mask entity information in the pattern and the map data Setting one piece of mask object information among a plurality of mask object information for each section including a cow; and generating data of a plurality of masks corresponding to the plurality of mask object information using the set mask object information. Include.

Other features of the present invention will become apparent from the following detailed description of exemplary embodiments with reference to the accompanying drawings.

1 is a flowchart showing a method of generating mask data according to the first exemplary embodiment of the present invention.
2A and 2B show patterns and map data, respectively.
3 shows a constraint region.
4A, 4B, and 4C show examples of setting mask numbers.
5 shows a conventional example in which polygon data is subjected to pattern division without being converted to map data.
6 shows an example of a pattern to be divided.
7 is a flowchart showing a method of generating mask data according to the second exemplary embodiment of the present invention.
8 shows an example of a conflict graph.

The present invention relates to semiconductor chips such as integrated circuits (ICs) and large-scale integration (LSI) circuits, display devices such as liquid crystal panels, detection elements such as magnetic heads, and charge-coupled devices. device: It is useful for the manufacture of various devices including image sensors such as CCD) sensors, and for generating data of patterns of masks (original plates) used in micromachines.

First, the mask data generation method according to the first exemplary embodiment will be described. 1 is a flowchart showing a method of generating mask data according to the first exemplary embodiment. This generating method is executed by an information processing apparatus such as a computer. According to this process, the data of the mask used for the exposure apparatus (lithographic apparatus) which exposes a board | substrate for transferring the image of a mask pattern to a board | substrate is produced | generated.

In step S101, the computer obtains data of a pattern to be divided. 2A shows a hole pattern or cut pattern used in a one-dimensional layout technique. In FIG. 2A, the pattern 2 comprises a plurality of pattern elements 20 each having the same rectangular shape. According to this embodiment, the pattern 2 is divided into patterns of a plurality of masks of, for example, a first mask, a second mask, and a third mask. In general, the data of the layout pattern used to form the pattern in the mask is converted and stored, for example, in the form of polygon data of GDSII. The computer acquires polygon data of the pattern 2.

In step S102, the computer converts the data of the pattern 2 in polygon data format into map data comprising a plurality of grids (or sections of any form) 21. An example of map data is shown in FIG. 2B. The area of the pattern 2 is divided into a plurality of squares (or sections of any shape) 21 having the same width so that each pattern element 20 is disposed in one of the sections 21. Then, in each square 21, "1" is assigned to the square containing the pattern element 20, and "0" is assigned to the square containing no pattern element 20. As shown in FIG. Since each pattern element 20 of the hole pattern or cut pattern is designed to fit the grid based on the grid design rule, each pattern element 20 has a one-to-one correspondence to each square 21 having the same width. Corresponding. In this manner, the data of the pattern 2 is converted into map data including information indicating the presence / absence of the pattern element 20 in each square 21. The shape of each square is not limited to a rectangle, and any shape may be used as long as each pattern element corresponds to each shape. Further, by matching the positions of the centers of the squares and the like with the intersections of the grids, grid map data in which the arrangement of each pattern element 20 is expressed by the coordinates of the respective intersections of the grids can be used as map data.

In step S103, the computer determines a constraint regarding the setting of the mask object information. A method of determining the constraint will be described with reference to FIG. 3. First, an area including one square 301 and squares 302 surrounding the square 301 is set as a constraint area. In FIG. 3, the constraint region is one square 301 (represented by ★) and squares 302 (represented by X) of up, down, left, right, inclined upwards and downwards with respect to the square 301. It includes. In this constraint area, since the squares are closely arranged, if the pattern elements of the squares are formed in one mask and subjected to exposure using the mask, the resolution of the pattern elements may be lowered. Therefore, a constraint condition is applied in this constraint area. According to this constraint, the use of the same mask object information for each square in the constraint area is prohibited.

The mask object information is, for example, number data or color data of the mask. The mask number is, for example, the individual number of each individual corresponding to each mask such as 1, 2, or 3. The squares of the constraint region are not limited to the upper, lower, left, right, inclined upper and lower squares with respect to the target square, and may include a square at any position at which resolution performance is degraded. For example, a square at a position separated by three squares from a square that is an evaluation target may be included in the constraint area.

In step S104, the computer selects one square as an evaluation target from the plurality of squares. In selecting a square, the computer selects a square comprising one of the pattern elements 21, which is a square whose value is "1". For example, the computer selects the square 401 shown in FIG. 4A. In step S105, the computer acquires mask object information that has already been set in the constraint area 402 around the square 401. In step S106, the computer sets mask object information different from mask object information already set in the constraint area 402 for the square 401 serving as the evaluation target.

A case where part of the mask object information (mask number) has already been determined will be described with reference to FIG. 4B. If the square 404 is a square of the evaluation target, the constraint region 405 will be a constraint region. Since the mask number "1" has already been used in the constraint area 405, the mask number "2" will be set in the square 404.

In step S107, the computer determines whether a square without mask object information exists among the plurality of squares including the pattern element 21. If there is a square without mask entity information (YES in step S107), the process returns to step S104 and will select a square as another evaluation target for which the mask entity information is to be set. Then, step S105 to step S107 are repeated. In step S104, a square with pattern element 21, which is a square having a value of "1", is selected, and an area set around the selected square is selected as the constraint area, but the selection of the square is not limited to this example. . For example, all squares may be sequentially selected regardless of the presence or absence of a pattern element, and the constraint area may be set by setting the selected square as the center. If no square exists without mask object information (NO in step S107), the process ends.

In this way, mask object information is set for all squares including the pattern element 21, and data is output. 4C shows map data with mask entity information set for all squares including pattern element 21. In Fig. 4C, a mask number of mask object information is set for each square.

Data of the mask pattern is generated according to the mask object information, and the pattern 2 is divided into a plurality of mask patterns according to the generated data. More specifically, the pattern 2 may include a first mask including pattern elements corresponding to mask number "1" and a second pattern including pattern elements corresponding to mask number "2", as shown in FIG. 4C. Two masks and patterns of a third mask including pattern elements corresponding to mask number "3".

There are several methods for setting mask object information in step S106. For example, there is a method of allocating the minimum number of mask numbers not used in the constraint area. For example, according to the example of FIG. 4B, when determining the mask number of the square 404 which is the evaluation target, since the mask number "1" has already been used in the constraint area 405, the minimum of the unused mask numbers is used. &Quot; 2 " Also, after determining the maximum mask number (at least 2), there is a method of randomly selecting one number below the maximum mask number among the mask numbers not used in the constraint area. For example, according to the example of Fig. 4B, when determining the mask number of the square 404 as the evaluation target, since the mask number "1" has already been used in the constraint region 405, the predetermined maximum mask number is "3". "," "2" or "3" is randomly determined as the mask number of the square 404. This method is referred to as the random digit method.

As conventionally performed, the case where the polygon data is divided into patterns without converting the polygon data into map data will be described with reference to FIG. First, as shown in FIG. 5, in order to determine the mask number of the pattern element 501 which is an evaluation target, the distance between the center of the pattern element 501 and the center of each of all the surrounding pattern elements is calculated. Then, a pattern element is determined that does not satisfy the constraint of the distance between the pattern elements. In addition, it is determined whether mask object information is already set in a pattern element that does not satisfy the constraint. If the mask object information is already set, mask object information different from the set mask object information will be set in the pattern element 501. The larger the number of pattern elements around the pattern element 501 that is the evaluation target, the longer it takes to calculate the distance.

On the other hand, according to the present embodiment, since the polygon data is converted into map data in which each pattern element is associated with each square, and the constraint area of the limited area is set based on the map data, Unnecessary calculations can be avoided and the calculation time can be reduced.

6 shows 5600 rectangular pattern elements. When the distance between the center positions of the pattern elements is calculated without converting the polygon data into the map data using a conventional method, and other mask object information is set for each square that does not satisfy the distance constraint. The calculation time was 4.99 seconds. On the other hand, when the polygon data was converted into map data and the mask number was set according to the random numbering method as performed in the present embodiment, the calculation time was 0.51 seconds. Thus, the calculation time was reduced to about 10%.

Next, a second exemplary embodiment of the present invention will be described. Fig. 7 is a flowchart showing a method of generating mask data according to the second exemplary embodiment.

In step S701, the computer acquires data of a pattern to be divided. In step S702, the computer converts data of the pattern of the polygon data format into map data partitioned into a plurality of squares. In step S703, the computer determines a constraint. Steps S701 to S703 are the same as steps S101 to S103 according to the first exemplary embodiment.

In step S704, the computer generates a conflict graph based on the map data. According to the conflict graph, the pattern elements are represented using points, and points exceeding the resolution limit are connected by sides. First, one square containing a pattern element is selected. Thereafter, in the constraint region including the selected square, the square including the pattern element and the selected square are connected by sides. This is done for all squares containing the pattern element. The mask object information at one end of the side of this graph needs to be different from the mask object information at the other end. In this manner, constraints on setting mask object information are set. FIG. 8 shows a conflict graph of the map data shown in FIG. 2B. It is obtained using the constraint area shown in FIG.

In step S705, the computer sets mask object information using integer programming. An example of integer programming aimed at minimizing the number of masks is represented by the following equation.

(1) Explanation of Variables

(2) Explanation of the equation

The cost function (the objective function) is expressed as

This means that the number of mask numbers to be used will be minimized.

(7)

(7)

Since Equation 7 is equal to the number of mask numbers, if the number of mask numbers increases from "2" to "3", the value of Equation 7 will change from "2" to "3". The number of mask numbers is the number of masks formed by dividing the pattern. Therefore, from the viewpoint of the cost of the mask, it is important to divide the pattern into fewer masks.

Constraints are represented by the equation

(8)

(8)

(9)

(9)

(10)

(10)

(11)

(11)

The boundary condition is represented by the following equation.

(12)

(12)

Equation 8 shows a constraint that the mask numbers need to be assigned in ascending order. The constraint is that if the first mask number is not used (y1 = 0), the use of the second mask number (y2 = 1) is prevented.

Equation 9 indicates that only one mask number is to be set for xi, the i-th pattern element. The constraint is that both the first mask number and the second mask number are prevented from being set for the i-th pattern element.

Equation 10 indicates that the mask number will not be set using an unused mask number. That is, when the j th mask number is not used (yj = 0), the constraint is to prevent the use of the j th mask number (xij = 1) for the i th pattern element.

Equation 11 represents a constraint of a pattern based on a conflict graph. This constraint is used when the i-th pattern and the i-th pattern are connected by line segments. That is, it is used when the same mask number should not be assigned. Thus, this constraint does not apply to all pattern elements. This constraint is set for the pattern elements in one constraint area.

For the boundary condition, the first mask number will be used as shown in equation (12).

In step S705, the computer inputs the above equations into the execution software for integer programming, performs calculations, and sets mask object information. The computer then outputs the data of the squares to which the mask entity information has been added.

According to this exemplary embodiment, when generating pattern data of a plurality of masks, the pattern is divided in a shorter time.

The above-described exemplary embodiments may be achieved by supplying a software program that implements the functions of each of the above-described exemplary embodiments to a system or apparatus via a network or various storage media, and may be implemented by a computer (or CPU) of the system or apparatus. Or MPU) reads and executes a program stored in such a storage medium.

Data of the mask generated as described above is input to the mask drawing apparatus, and a plurality of masks are manufactured. The manufactured mask is mounted on the mask stage of the exposure apparatus and is illuminated by the illumination optical system. In accordance with this illumination, the image of the mask pattern is exposed on the wafer. After exposing the wafer using one of the manufactured masks, the same layer of wafer is exposed using another mask. By doing so, a pattern can be formed on one layer of the wafer by multiple patterning.

Next, the manufacturing method of devices, such as a liquid crystal display device, is demonstrated. The manufacturing process for a liquid crystal display device includes the process of forming a transparent electrode. The process of forming a transparent electrode applies a photosensitive agent to the glass substrate in which the transparent conductive coating was deposited, sets the mask manufactured as mentioned above to an exposure apparatus, exposes the glass substrate to which the photosensitive agent was applied, and develops a glass substrate. The steps include.

In the above-described exposure apparatus, the illumination optical system illuminates the mask with light from the light source, and transfers the circuit pattern formed on the mask onto the wafer through the projection optical system or the like. However, exemplary embodiments are not limited to such exposure apparatus. The exposure apparatus may be an imprint apparatus configured to form a pattern on a substrate by contacting a pattern formed on a mold with a resin applied on the substrate, and then curing or curing the resin. have. In this case, the data of the divided mask pattern corresponds to the pattern formed on the mold. In addition, the exposure apparatus described in the exemplary embodiments may be an apparatus configured to draw on the substrate by a charge particle beam (electron beam). In this case, the apparatus configured to perform drawing is controlled based on the data of the divided mask pattern. Further, images of the patterns can be formed by using a combination of the various devices described above for each divided mask pattern.

The device manufacturing method mentioned above using an exposure apparatus is suitable also for devices, such as a semiconductor device, besides manufacture of a liquid crystal display device. The method may include mounting a mask manufactured as described above in an exposure apparatus, exposing a substrate to which the photosensitive agent is applied, and developing the illuminated substrate. The device manufacturing method may also include other well known processes such as oxidation, film formation, deposition, doping, planarization, etching, resist stripping, dicing, bonding, and packaging.

Embodiments of the present invention read and execute computer executable instructions recorded on a storage medium (eg, a non-transitory computer readable storage medium) to perform the functions of one or more of the above-described embodiment (s) of the present invention. May be implemented by a computer of a system or apparatus, and for example, by reading and executing computer executable instructions from a storage medium to perform the functions of one or more embodiment (s) described above. It may be implemented by a method performed by a computer. The computer may include one or more of a central processing unit (CPU), micro processing unit (MPU), or other circuitry, and may include separate computers or a network of separate computer processors. Can be. The computer executable instructions may be provided to the computer, for example, from a network or the storage medium. Storage media may include, for example, hard disks, random access memory (RAM), read-only memory (ROM), storage of distributed computing systems, optical disks (compact disks (CD), digital versatile disks (DVD), or blue-blue). Ray disk (BD) ^™, etc.), flash memory devices, memory cards, and the like.

Although the present invention has been described with reference to exemplary embodiments, it will be understood that the invention is not limited to the disclosed exemplary embodiments. The scope of the following claims is to be accorded the broadest interpretation so as to encompass all such modifications and equivalent structures and functions.

Claims (13)

After the patterning of the substrate using the first mask, the data of the plurality of masks including the first mask and the second mask, which are used for multiple patterning of patterning the substrate using the second mask, using a computer Mask data generation method to generate the
Acquiring data of a pattern including a plurality of pattern elements,
A map including information indicating the presence or absence of the pattern element in each section by dividing an area of the pattern into a plurality of sections so that each pattern element is disposed in each section using the obtained data of the pattern Generating data,
A constraint for prohibiting setting of the same mask entity information in a constraint region including one section and sections surrounding the section, and using the map data, a plurality of masks for each section including the pattern element Setting mask object information of one of the object information, and
And generating data of a plurality of masks corresponding to the plurality of mask object information by using the set mask object information. The method of claim 1,
And the constraint region is an area comprising one section and surrounding sections surrounding the one section. The method of claim 1,
And the constraint area is an area including one section and sections neighboring the one section. The method of claim 1,
The setting step,
Selecting one of the plurality of sections, and setting mask object information different from mask object information set in the selected section to a section including the pattern element in the constraint region including the selected section, Mask data generation method for setting a plurality of mask object information by repeating for the plurality of sections. The method of claim 1,
The setting step,
Mask data generation method which sets the mask object information randomly selected from the said mask object information which was not used in the said constraint area among the said plurality of mask object information. The method of claim 1,
The setting step,
Mask data generation method for setting the mask object information using integer programming. The method according to claim 6,
The integer programming uses the number of mask entity information as a cost function and sets the mask entity information to minimize the cost function. The method according to claim 6,
When setting a section including the pattern element among the plurality of sections as a point, the setting step includes setting a constraint by connecting points to sides in the constraint area where setting of the same mask object information is prohibited. And setting the mask object information using the integer programming so as not to set the same mask object information in the section of the points connected by the sides. The method of claim 1,
And the mask object information is number data or color data. A computer-readable storage medium storing a program for executing a method of generating mask data according to any one of claims 1 to 9 in a computer. An information processing apparatus that executes the mask data generation method according to any one of claims 1 to 9. A lithographic apparatus for patterning a substrate using a pattern formed in the first mask and a pattern formed in the second mask in multiple patterning of patterning the substrate using a second mask after patterning the substrate using a first mask. ,
Data of a plurality of masks including the first mask and the second mask is generated by a mask data generation method,
The mask data generation method is a method of generating data of a plurality of masks using a computer,
Acquiring data of a pattern including a plurality of pattern elements,
A map including information indicating the presence or absence of the pattern element in each section by dividing an area of the pattern into a plurality of sections so that each pattern element is disposed in each section using the obtained data of the pattern Generating data,
One of a plurality of mask object information for each section including the pattern element, using a constraint for prohibiting setting of the same mask object information in a constraint area including one section and sections surrounding it, and the map data Setting mask object information for, and
And generating data of a plurality of masks corresponding to the plurality of mask object information using the set mask object information. A device manufacturing method comprising patterning a substrate using a lithographic apparatus, and developing the patterned substrate,
The lithographic apparatus is configured to pattern a substrate using a pattern formed on the first mask and a pattern formed on the second mask in multiple patterning of patterning the substrate using a second mask after patterning the substrate using a first mask. Patterning,
Data of a plurality of masks including the first mask and the second mask is generated by a mask data generation method,
The mask data generation method is a method of generating data of a plurality of masks using a computer,
Acquiring data of a pattern including a plurality of pattern elements,
A map including information indicating the presence or absence of the pattern element in each section by dividing an area of the pattern into a plurality of sections so that each pattern element is disposed in each section using the obtained data of the pattern Generating data,
One of a plurality of mask object information for each section including the pattern element, using a constraint for prohibiting setting of the same mask object information in a constraint area including one section and sections surrounding it, and the map data Setting mask object information for, and
And generating data of a plurality of masks corresponding to the plurality of mask object information using the set mask object information.

Images